Thanks! It was born out of complete stubbornness 
3 Likes
I spent a lot of time in CAD trying to figure out if it would work at all. Originally I had the bottom jig for the groups the same on all of them, but I realized the enclosure tapers in a bit towards the bottom and it made them super tight when I put them in that way. I redesigned the jigs for the side groups to have a bevel on the corner so they could fit with squeezing against each other (they all had plenty of wiggle room).
They were all like that originally ^ before I realized I needed the bevel, and that I’d need to use the empty channel on top for wires. I could run it like that but I’d need a thick gasket for the enclosure and it’s already deep enough as it is.
You can see the beveled corners on the underside there from when I was testing it heat shrinked (shrunk? 
) to make sure it still fit the enclosure ok.
5 Likes
I haven’t seen this done yet. Very nice job.
3 Likes
Creating a significant bottleneck with your nickel with this method.
1 Like
Yeah, it’s less than ideal. If I had another 6-7mm on top of the pack I would have done my original method and folded a larger piece over the top. This thing is maxed out on available space for sure.
This is how I wanted to do it originally when mocking it up with printed P groups but with the wires soldered on it would have been smushed against the deck and that’s no bueno
Yes technically, but with the nickel folded on one of the top angles it made a sharp point right at the top that I was not a big fan of. I could made the tab smaller to avoid the point, but it was essentially the same size as it is now (around 12mm wide)
1 Like
Yeah but you could have had one tab on each side
Anyone ever had this happen?
Some groups fully disappeared, so I checked all connections on the BMS, all were fine. Then I just plugged it into a charger and suddenly the dead groups reappeared…
Don’t like this at all!
1 Like
Im so paranoid about bms, that is no Bueno. Is that the Xiaoxiangbms app? Maybe it is an app error but i have never seen that. Definitely if it does anything weird again replace it and keep an eye on it closely while charging and order a new bms anyway just because.
This is an alternative to the xiaoxiang app called Overkill solar.
The xiaoxiang app showed the same data though.
Probably gonna replace it
2 Likes
It just seems odd that it wasn’t throwing an error or something. All your balance connections are glued or taped in place? Could one of the balance jst connectors have been loose but you didn’t notice ? That’s all i can think of, it does look like everything on one was missing but im not certain. Could test it and silicone or hot glue them down again just to make sure
Edit, nm that’s 12s right? One connector I believe
One, and it’s in perfect condition. All wires measure the correct voltage
1 Like
Hopefully someone with more experience can chime in here with the easy solution because I’ve not see this type of issue before. If it was mine I would order a new bms and shit can that one - no second chances from me anymore. I’ve done my fucking around and my tolerance for finding out is low.
If the app doesn’t see the pgroup then id assume the bms doesn’t see it. I would monitor charging very very closely and make sure it’s balancing only when charging so it doesn’t kill a p group to try and balance it to match one of the ghost groups.
I have a spare bms if you’re interested in not waiting for a slow boat but it’s not an exact match - it’s the 10? To 17s llt smart bms. Just whatever I paid and shipping unless you’re in Seattle area.
The fruit is not included, wont make customs inspection for Belgium
My first battery build with a Kweld! 
Any feedback would be appreciated.
This 10s4p pack with cheap BAK cells was a practice battery for my wifes low powered 250w ebike. The battery will sit inside a frame bag with foam cushioning.
- I used 0.2mm nickel plated steel. The wires were overkill for 25a max amps, but oh well.
- I used a 40a daly BMS with on/off switch
- Kweld settings: 25J. Did a test weld, and the nickel ripped leaving dots.
Step 1: Hot glued all the P groups
Step 2: Added fishpaper dots to separate the S groups.
Step 3: Hot glue all the S groups
Step 4: Spot weld!
Step 5: Add BMS leads
Step 6: Add the positive and negative leads
Step 7: Kapton tape, fish paper, heat shrink
It took me 4hrs from start to finish. And I think it turned out really well for a low powered practice battery
link, this battery got it’s own thread:
1 Like
Not bad. Here is few alarming things I spotted in your images:
You can see all this sharp edges on the positive side ? I marked them with red. I always use dremel to make them round and smooth and usually fit precisely with positive terminal (no overlaps). It’s probably harder when you use 2 pieces of nickle.
Here I marked in pink your balance wires. They are bending directly on the sharp nickle. Try to wire them to not touch nickel but if you have to put fish paper under it, few layers. You also don’t need such a thick wires.
This is a THICK wire but thin nickle connection 
Your heat will concentrate in the middle where 3 pink arrows touches. You should distribute it. Best way is to solder wire directly to nickle like I painted in green. This solder connection does not look encouraging but maybe it’s not that bad.
This I think is the biggest red flag. You need to put fish paper when green arrow is pointing.
4 Likes
thanks for the feedback. I’ll keep them in mind for the next build. And I’ll add fishpaper by the positive lead exit area
Oh shit. I didn’t see those sharp nickel corners on the positive cells.
If the deck you’re putting this on is flexy at all you really need to go back in and snip all of those. A quality pair of knippers will make short work of them.
But those will 100% short the pack & likely start a fire if this battery sees a decent amount of usage.
1 Like
Thx for the tip. Set my DV4s 45 A per side and the thing is like a rocket on dual 6374 motors
1 Like